Observation of Attention Mechanism Baseline for PCB Surface Inspection System

Fityanul Akhyar; Ledya Novamizanti; Muhammad Azka Imaddudin; Ikhsanico Henda Pratama; Shandy Ramanda Firmansyach; Ming Ching Chang; Chih Yang Lin

doi:10.1109/APWiMob56856.2022.10014223

Observation of Attention Mechanism Baseline for PCB Surface Inspection System

Fityanul Akhyar, Ledya Novamizanti, Muhammad Azka Imaddudin, Ikhsanico Henda Pratama, Shandy Ramanda Firmansyach, Ming Ching Chang, Chih Yang Lin

Department of Mechanical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

Printed circuit boards (PCBs) are critical for interconnecting various components and allowing them to communicate with each other. It is critical to ensure that there are no small surface defects that can negatively impact PCB production. Therefore, template matching is often used in PCB surface inspection systems. Despite its popularity, this method can be improved because inspecting a PCB with a template is inefficient. Currently, integrating the surface inspection system with the deep learning method is proving to be more effective in solving this problem. This paper examines three popular deep learning object recognition methods in order to determine which one is the most effective in terms of attention. These three models are called Carafe, Empirical Attention, and ResNeSt. The experimental results showed that ResNeSt with split attention networks achieves the greatest accuracy in deep learning PCB surface inspection system with a mean average precision (mAP) of 99.2% and an average recall (AR) of 99.5%. The result of this study would improve the effectiveness of PCB surface inspection in controlling production lines.

Original language	English
Title of host publication	APWiMob 2022 - Proceedings
Subtitle of host publication	2022 IEEE Asia Pacific Conference on Wireless and Mobile
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781665474863
DOIs	https://doi.org/10.1109/APWiMob56856.2022.10014223
State	Published - 2022
Event	2022 IEEE Asia Pacific Conference on Wireless and Mobile, APWiMob 2022 - Bandung, Indonesia Duration: 9 Dec 2022 → 10 Dec 2022

Publication series

Name	APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile

Conference

Conference	2022 IEEE Asia Pacific Conference on Wireless and Mobile, APWiMob 2022
Country/Territory	Indonesia
City	Bandung
Period	9/12/22 → 10/12/22

Keywords

PCB
attention network
deep learning
defect detection

Access to Document

10.1109/APWiMob56856.2022.10014223

Cite this

Akhyar, F., Novamizanti, L., Imaddudin, M. A., Henda Pratama, I., Firmansyach, S. R., Chang, M. C., & Lin, C. Y. (2022). Observation of Attention Mechanism Baseline for PCB Surface Inspection System. In APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/APWiMob56856.2022.10014223

Akhyar, Fityanul ; Novamizanti, Ledya ; Imaddudin, Muhammad Azka et al. / Observation of Attention Mechanism Baseline for PCB Surface Inspection System. APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile. Institute of Electrical and Electronics Engineers Inc., 2022. (APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile).

@inproceedings{4b7ebe5bd4004f05b3c7177866322272,

title = "Observation of Attention Mechanism Baseline for PCB Surface Inspection System",

abstract = "Printed circuit boards (PCBs) are critical for interconnecting various components and allowing them to communicate with each other. It is critical to ensure that there are no small surface defects that can negatively impact PCB production. Therefore, template matching is often used in PCB surface inspection systems. Despite its popularity, this method can be improved because inspecting a PCB with a template is inefficient. Currently, integrating the surface inspection system with the deep learning method is proving to be more effective in solving this problem. This paper examines three popular deep learning object recognition methods in order to determine which one is the most effective in terms of attention. These three models are called Carafe, Empirical Attention, and ResNeSt. The experimental results showed that ResNeSt with split attention networks achieves the greatest accuracy in deep learning PCB surface inspection system with a mean average precision (mAP) of 99.2% and an average recall (AR) of 99.5%. The result of this study would improve the effectiveness of PCB surface inspection in controlling production lines.",

keywords = "PCB, attention network, deep learning, defect detection",

author = "Fityanul Akhyar and Ledya Novamizanti and Imaddudin, {Muhammad Azka} and {Henda Pratama}, Ikhsanico and Firmansyach, {Shandy Ramanda} and Chang, {Ming Ching} and Lin, {Chih Yang}",

note = "Publisher Copyright: {\textcopyright} 2022 IEEE.; 2022 IEEE Asia Pacific Conference on Wireless and Mobile, APWiMob 2022 ; Conference date: 09-12-2022 Through 10-12-2022",

year = "2022",

doi = "10.1109/APWiMob56856.2022.10014223",

language = "???core.languages.en_GB???",

series = "APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "APWiMob 2022 - Proceedings",

}

Akhyar, F, Novamizanti, L, Imaddudin, MA, Henda Pratama, I, Firmansyach, SR, Chang, MC & Lin, CY 2022, Observation of Attention Mechanism Baseline for PCB Surface Inspection System. in APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile. APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile, Institute of Electrical and Electronics Engineers Inc., 2022 IEEE Asia Pacific Conference on Wireless and Mobile, APWiMob 2022, Bandung, Indonesia, 9/12/22. https://doi.org/10.1109/APWiMob56856.2022.10014223

Observation of Attention Mechanism Baseline for PCB Surface Inspection System. / Akhyar, Fityanul; Novamizanti, Ledya; Imaddudin, Muhammad Azka et al.
APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile. Institute of Electrical and Electronics Engineers Inc., 2022. (APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Observation of Attention Mechanism Baseline for PCB Surface Inspection System

AU - Akhyar, Fityanul

AU - Novamizanti, Ledya

AU - Imaddudin, Muhammad Azka

AU - Henda Pratama, Ikhsanico

AU - Firmansyach, Shandy Ramanda

AU - Chang, Ming Ching

AU - Lin, Chih Yang

PY - 2022

Y1 - 2022

N2 - Printed circuit boards (PCBs) are critical for interconnecting various components and allowing them to communicate with each other. It is critical to ensure that there are no small surface defects that can negatively impact PCB production. Therefore, template matching is often used in PCB surface inspection systems. Despite its popularity, this method can be improved because inspecting a PCB with a template is inefficient. Currently, integrating the surface inspection system with the deep learning method is proving to be more effective in solving this problem. This paper examines three popular deep learning object recognition methods in order to determine which one is the most effective in terms of attention. These three models are called Carafe, Empirical Attention, and ResNeSt. The experimental results showed that ResNeSt with split attention networks achieves the greatest accuracy in deep learning PCB surface inspection system with a mean average precision (mAP) of 99.2% and an average recall (AR) of 99.5%. The result of this study would improve the effectiveness of PCB surface inspection in controlling production lines.

AB - Printed circuit boards (PCBs) are critical for interconnecting various components and allowing them to communicate with each other. It is critical to ensure that there are no small surface defects that can negatively impact PCB production. Therefore, template matching is often used in PCB surface inspection systems. Despite its popularity, this method can be improved because inspecting a PCB with a template is inefficient. Currently, integrating the surface inspection system with the deep learning method is proving to be more effective in solving this problem. This paper examines three popular deep learning object recognition methods in order to determine which one is the most effective in terms of attention. These three models are called Carafe, Empirical Attention, and ResNeSt. The experimental results showed that ResNeSt with split attention networks achieves the greatest accuracy in deep learning PCB surface inspection system with a mean average precision (mAP) of 99.2% and an average recall (AR) of 99.5%. The result of this study would improve the effectiveness of PCB surface inspection in controlling production lines.

KW - PCB

KW - attention network

KW - deep learning

KW - defect detection

UR - http://www.scopus.com/inward/record.url?scp=85147539010&partnerID=8YFLogxK

U2 - 10.1109/APWiMob56856.2022.10014223

DO - 10.1109/APWiMob56856.2022.10014223

M3 - 會議論文篇章

AN - SCOPUS:85147539010

T3 - APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile

BT - APWiMob 2022 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2022 IEEE Asia Pacific Conference on Wireless and Mobile, APWiMob 2022

Y2 - 9 December 2022 through 10 December 2022

ER -

Akhyar F, Novamizanti L, Imaddudin MA, Henda Pratama I, Firmansyach SR, Chang MC et al. Observation of Attention Mechanism Baseline for PCB Surface Inspection System. In APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile. Institute of Electrical and Electronics Engineers Inc. 2022. (APWiMob 2022 - Proceedings: 2022 IEEE Asia Pacific Conference on Wireless and Mobile). doi: 10.1109/APWiMob56856.2022.10014223

Observation of Attention Mechanism Baseline for PCB Surface Inspection System

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this